co-digestion of source segregated domestic food waste to improve ...€¦ · domestic food waste to...
TRANSCRIPT
1
This is a revised personal version of the text of the final journal article, which is made available for scholarly purposes only, in accordance with the journal's author permissions. The full citation is: Zhang Y., Banks C. J., Heaven S. (2012) Co-digestion of source segregated domestic food waste to improve process stability. Bioresource Technology 114, 168-178. doi: 10.1016/j.biortech.2012.03.040 ___________________________________________________________________________________________________
Co-digestion of source segregated domestic food waste to improve process stability
Author names and affiliations
Yue Zhang1, Charles J. Banks, Sonia Heaven
Faculty of Engineering and the Environment, University of Southampton, Southampton
SO17 1BJ, UK
Abstract
Cattle slurry and card packaging were used to improve the operational stability of food
waste digestion, with the aim of reducing digestate total ammoniacal nitrogen
concentrations compared to food waste only. Use of cattle slurry could have major
environmental benefits through reducing greenhouse gas emissions associated with
current management practices; whilst card packaging is closely linked to food waste and
could be co-collected as a source segregated material. Both options increase the
renewable energy potential whilst retaining organic matter and nutrients for soil
replenishment. Co-digestion allowed higher organic loadings and gave a more stable
process. A high ammonia inoculum acclimated more readily to cattle slurry than card
1 * Corresponding author: Tel.: +44 (0)2380 598363; fax: +44 (0)2380 677519; E-mail address:
[email protected] (Yue Zhang)
2
packaging, probably through supplementation by trace elements and micro-organisms.
Long-term operation at a 75-litre scale showed a characteristic pattern of volatile fatty
acid accumulation in mono-digestion of food waste, and allowed performance
parameters to be determined for the co-digestion substrates.
Keywords: Food waste; card packaging; cattle slurry; ammonia; specific methane
production
1. Introduction
Anaerobic digestion of organic material in municipal solid waste provides renewable
energy in the form of biogas (Mata Alvarez, 2003) and can also offer a means of
recycling valuable plant nutrients back to agricultural land (Lukehurst et al., 2010). To
achieve the latter requires source segregation of targeted organics, as the material
arising from mechanical pre-treatment has a high level of contamination, with heavy
metal concentrations exceeding the accepted values for agricultural land used in food
production (BioAbV, 1998; PAS 110, 2010). Food waste from both domestic and
commercial sources has been targeted for biogas production because of its high
biochemical methane potential (Zhang et al., 2011; Banks and Zhang, 2010), whilst its
high water content makes energy recovery through thermal treatment unattractive
(Ahring, 2003). It can, however, be difficult to digest as a mono-substrate (Zhang et al.,
2011), leading to digester instability and in some cases failure (Neiva Correia et al.,
2008; Resch, 2011; Palatsi, 2011). Recent work by Banks et al. (2012) has shown that
stable digestion is possible at the high ammonia concentrations associated with food
waste by selective trace element addition. An alternative approach is to co-digest food
3
waste with other waste materials so as to increase the carbon to nitrogen ratio as a
means of overcoming process limitations due to ammonia inhibition (Zhang et al.,
2011). There are also strong environmental reasons for adopting co-digestion: treating
animal manures in a controlled process reduces the fugitive emissions associated with
manure management and could lead to greenhouse gas savings (Clemens, 2006; Banks
et al., 2007; Marañón et al., 2011). The mixing of a high energy potential substrate such
as food waste with low energy potential animal slurries can make the overall process
economic (Angelidaki and Ellegaard, 2003; El-Mashad and Zhang, 2010; Zhang et al.,
2011). A number of studies have explored this option and shown improved performance
or increased process stability (Callaghan et al., 2002; Hartmann and Ahring, 2005;
Capela et al., 2007; Alvarez and Lidén, 2008). The concept has also been successfully
applied to achieve better nutrient management by cooperative schemes in Denmark
(Holm-Nielsen et al., 2009), mainly using commercial or industrial sources of
biodegradable wastes from animal slaughter and food processing (Raven and Gregersen,
2007). These schemes are often regarded as a model of best practice (Braun and
Wellinger, 2003); not all of the Danish co-digestion plants have worked without
problems, however, probably due to unwise selection of co-substrates (Nielsen and
Angelidaki, 2008).
Banks et al. (2011a) suggested that on-farm co-digestion of source segregated domestic
food waste was the most effective means of making cattle slurry digestion economically
viable, with associated benefits in greenhouse gas reduction and nutrient management.
In dense urban areas where centralised digestion may be more appropriate, card
packaging material becomes an attractive co-substrate as it is generated and can be co-
4
collected in close association with food waste, and is generally available in tonnage
quantities as a low or negative value stream from materials recovery facilities. The aim
of the current work was to investigate the co-digestion of source segregated domestic
food waste with cattle slurry and also with card packaging: in both cases it was
anticipated that the carbon/nitrogen (C/N) ratio would be reduced, allowing operation at
higher loading rates with enhanced process stability.
2. Methods
2.1. Materials
2.1.1. Digesters
The work was carried out in laboratory and pilot-scale mechanically mixed mesophilic
digesters which were fed daily. The laboratory-scale digesters each had a 4-litre working
volume and were constructed of PVC tube with gas-tight top and bottom plates. The top
plate was fitted with a gas outlet, a feed port sealed with a rubber bung, and a draught tube
liquid seal through which a stainless steel asymmetric bar stirrer was inserted with a 40
rpm motor mounted directly on the top plate. Temperature was maintained at 36 ± 1 °C
by circulating water from a thermostatically-controlled bath through a heating coil around
the digesters. Semi-continuous operation was achieved by daily removal of digestate
through an outlet port in the base of each digester, followed by substrate addition via the
feed port. The pilot-scale digesters were of the same design as the laboratory-scale
digesters except that their working volume was 75 l and the temperature was maintained
at 36 ± 1 oC by an internal heat exchanger. In both cases biogas production was
5
measured using tipping-bucket gas counters with continuous data logging (Walker et al.,
2009).
2.1.2. Food waste
Source segregated domestic food waste (FW) delivered to the South Shropshire digestion
facility at Ludlow, UK was used in the study (Banks et al., 2011b). The material was first
taken out of biodegradable plastic bags and any non-biodegradable contaminants were
removed. It was then homogenised using a macerating grinder (S52/010 Waste Disposer,
IMC Limited, UK), packed into 4-litre plastic storage containers, and frozen at −18 oC.
Before use the feedstock was thawed, and stored at 4 oC for no more than one week. The
characteristics of the FW are given in Table 1.
2.1.3. Cattle slurry
Three batches of cattle slurry (CS) was obtained from Parker's Farm, Hampshire, UK,
over the period of experiment. It was then homogenised and stored in the same way as for
food waste. Table 1 shows the typical characteristics of the material.
2.1.4. Card packaging
One hundred kilograms of mixed card packaging (CP) was collected from the reject
stream of the Alton Materials Recovery Facility (MRF) run by Veolia Environmental
Services, Hampshire, UK. This material was sorted into three fractions and shredded
into 50 × 5 mm pieces using office shredders, and then blended again in defined
proportions consisting of 29.6% of corrugated cardboard, 62.5% of card packaging and
7.9% of other card on a fresh weight basis. The proportions were calculated based on
6
these different recyclable carton types in the UK household waste stream. Table 1 gives
the characteristics of the material.
2.1.5. Inoculum
The inoculum for the 4-litre laboratory-scale trial was taken from a 35-litre food waste
digester that had been acclimated to this substrate over a period of 284 days at an
organic loading rate (OLR) of 2 kg volatile solids (VS) m−3
day−1
(Banks and Zhang,
2010). Before use the digestate was sieved through a 1 mm mesh to remove any large
particles. The 75-litre digesters were inoculated with digestate from a mesophilic
digester treating municipal wastewater biosolids at Millbrook Wastewater Treatment
Plant (WWTP), Southampton, UK.
2.2. Digester operation and monitoring
Two laboratory-scale digesters were fed with a mixture of FW and CS, at an initial ratio
of 20:80% on a VS basis. Two more digesters were fed with FW and CP in the
proportion of 78.4:21.6 on a fresh weight basis which was calculated based on their
relative quantities in the UK household waste stream, giving a FW:CP ratio of 53:47 on
a VS basis. To ensure homogeneity the card packaging was prepared by wet maceration,
which reduced the total solids content from 94% to around 20%. A fifth digester was
fed only on FW and acted as a control. The total OLR for all digesters was 2 kg VS m−3
day−1
at the start of the trial. The solids and liquid retention times were uncoupled and a
nominal solids retention time (SRT) of 30 days was maintained by sieving digestate
through a 1 mm mesh, after which the appropriate quantity of solids was discarded and
the amount of liquid required to maintain the digester at a constant volume was returned
7
to the digester with the fresh substrate. The operating regime changed as the experiment
progressed, and details are given in Table 2. The trial ran for 329 days during which
time the loading was successively increased, and the proportion of FW:CS was also
raised to a final value of 60:40 on a VS basis to cover a range of operating conditions
with different proportions of the co-substrates.
The OLR in the 75-litre digesters was 2 kg VS m−3
day−1
and a nominal SRT of 30 days
was maintained through liquor re-circulation. One digester was fed on FW only, one on
a FW:CS mix at a 20:80 ratio on a VS basis, corresponding to the initial mixture in the
laboratory-scale trials; and one on a FW:CP mix at a 53:47 VS ratio. This ratio was the
same as used in the laboratory-scale trial, but in this trial the card packaging was not
wet macerated. The digesters were operated for 308 days, or more than 10 solids
retention times.
All digesters were monitored daily for biogas production and pH. Other digestate
parameters such as solids, volatile fatty acids (VFA), total ammoniacal nitrogen (TAN),
alkalinity, and biogas composition were analysed a minimum of once per week and
often more frequently. All gas volumes reported are corrected to standard temperature
and pressure (STP) of 0 oC, 101.325 kPa.
2.3. Digestate biostability
Digestate stability assays were carried out in 1.4-litre working capacity continuously
stirred tank reactor (CSTR) digesters at 36 ± 1 °C. The biogas produced was collected
using gas impermeable sampling bags (SKC Ltd., UK), with biogas volumes measured
8
according to the weight gasometer method (Walker et al., 2009). Inoculum was taken
from the Millbrook WWTP digester and sieved through a 1 mm mesh before use.
Whole digestate and digestate fibre were tested at an inoculum-to-substrate (i/s) ratio of
2 on a VS basis. In the case of digestate liquors, no inoculum was added and the
residual gas production occurred as a result of the microbial consortium already present.
The test was set up immediately after digestate had been drained from the digester and
the liquor separated from the fibre, and ran for 100 days. The test materials were run in
duplicate, with duplicate inoculum-only blanks and positive controls.
2.4. Analytical methods
Total solids (TS) and volatile solids (VS) were measured using Standard Method 2540
G (APHA, 2005). pH was determined using a Jenway 3010 meter (Bibby Scientific
Ltd., UK) with a combination glass electrode, calibrated in buffers at pH 4.0, 7.0 and
9.2 (Fisher Scientific, UK). Alkalinity was measured by titration with 0.25 N H2SO4 to
endpoints of pH 5.75 and 4.3, allowing calculation of total (TA), partial (PA) and
intermediate alkalinity (IA) (Ripley et al., 1986). Total Kjeldahl nitrogen (TKN) was
determined using a Kjeltech block digester and TAN by steam distillation unit
according to the manufacturer's instructions (Foss Ltd., Warrington, UK). Volatile fatty
acids were quantified in a Shimazdu GC-2010 gas chromatograph (Shimadzu, Milton
Keynes, UK), using a flame ionisation detector and a capillary column type SGE BP-21.
Biogas composition (CH4 and CO2) was determined using a Varian star 3400 CX gas
chromatograph, calibrated with 65% (v/v) CH4 and 35% (v/v) CO2. Trace element
concentrations were determined using ICP−MS or ICP−OES at a commercial laboratory
9
Severn Trent Services (Coventry, UK) after in-house hydrochloric−nitric acid digestion
(SCA, 1986).
3 Results and discussion
3.1. 4-litre digester trial
The inoculum used in this trial was deliberately drawn from a digester previously fed on
the same type of food waste at a low OLR over a long period. The aim was to show that
continued feeding of this well-established but already stressed inoculum would lead to
further deterioration, and to allow comparison of the effect of co-substrate addition.
3.1.1. Food waste only
The inoculum used had an initial total VFA concentration above 7000 mg l−1
, TAN
close to 3500 mg l−1
and free ammonia around 270 mg l−1
. Within one SRT, the specific
methane production (SMP) of the digester fed with FW as a single substrate fell to less
than 10% of its starting value. The pH had dropped below 6.0, the biogas contained less
than 10% methane and a high hydrogen peak was seen in GC analysis. These results
indicated that the digestion had failed and feeding was stopped, but the biogas
production, biogas composition and digestate parameters were still monitored.
3.1.2. Food waste and cattle slurry
In the two digesters fed with the FW:CS mix, stable biogas production was achieved
within about 0.5 SRT. Measured over the period day 30−98 the specific methane
production was 0.218 STP m3 CH4 kg
−1 VSadded (Fig. 1a); as expected, this was lower
than the value obtained for digestion of FW only at the beginning of the trial. The
10
biogas methane content (Fig. 1b) increased and reached a value of 62.7%. There was
also a decrease in the VFA concentration in the digesters (Fig. 1c and d), which may
have been due in part to methane production, and in part to hydraulic flushing of VFA
out of the system. The propionic acid (HPr) concentration decreased rapidly, coupled
with a rise in acetic acid (HAc). At the end of the first SRT HPr formed only 3.8% of
total VFA with HAc being the predominant species at 92%. After 1.5 SRT (day 45)
HAc was the only VFA species present in the digestate, and after 2 SRT (day 60) the
total VFA concentration had dropped to less than 200 mg l−1
. These results indicated
that the addition of CS had stabilised the digestion and had also brought about a
reduction in total ammoniacal nitrogen (Fig. 1e) and a stable pH between 7.5 and 8.0
(Fig. 1f).
The digesters running on a FW:CS mix were considered to have reached steady state
after 3 SRT, and on day 99 the proportion of FW in the mixture was increased to 40%
on a VS basis to improve the specific and volumetric methane production (Fig. 1a). The
digesters adapted to this change without VFA accumulation (Fig. 1c and d), and ran
stably for a further period of 2.5 SRT. From Fig. 1a it can be seen that the specific
methane production increased by about 10% and the volumetric methane production
increased from 0.43 to 0.48 STP m3
CH4 m−3
day−1
.
At the end of the sixth SRT (day 176) the OLR was increased to 3 kg VS m−3
day−1
with
the proportion of FW:CS remaining at 40:60. In one of the digesters this caused a rapid
increase in VFA concentration followed by a fall (Fig. 1c), coupled with a fluctuating
methane content and yield (Fig. 1a and b). Unstable biogas and methane production was
11
also observed in the other digester, although VFA concentrations remained below 200
mg l−1
(Fig. 1d); it may be that a VFA spike in this digester was not seen, if it
disappeared rapidly between sampling times. In Fig. 1a it can be seen that after the OLR
rose to 3 kg VS m−3
day−1
the volumetric methane production increased by 38% to reach
0.66 STP m3 CH4 m
−3 day
−1, but the specific methane production declined to 92% of
that obtained at an OLR of 2 kg VS m−3
day−1
. The probable reason for this is because
the cattle slurry had a relatively low solids content (only around 7% of VS in fresh
matter), and therefore the nominal liquid retention time was reduced to around 35 days,
causing a higher proportion of undegraded material to be flushed out of the digesters.
On day 232, after the digesters had run at an OLR of 3 kg VS m−3
day−1
for 2 SRT, the
proportion of food waste in the mixture was increased from 40% to 60%. This reduced
the volume of feedstock needed to achieve the required OLR, mitigating the wash-out
effect caused by the low solids content of cattle slurry. In this way it was possible to
obtain a further increase in OLR without reducing the solid and liquid retention times to
less than 30 days. The digesters adapted smoothly to the change in feedstock
proportions, and the volumetric methane production increased by 34% compared to that
for a 40:60 FW:CS mix (Fig. 1a). All stability parameters remained in a safe range with
VFA concentrations <150 mg l−1
, TAN 1600 mg l−1
, and a pH of 7.5.
To further improve the volumetric productivity, at the end of the tenth SRT (day 295)
the OLR was increased to 4 kg VS m−3
day−1
, with the proportion of food waste to cattle
slurry remaining at 60:40 on a VS basis. At the end of the trial the digesters had been
running under this regime for more than one SRT without any fluctuation in stability
12
parameters. This is longer than necessary for any loading shock to become apparent,
and it was therefore considered that the operation of the digesters corresponded to
steady state conditions with a specific methane production of 0.306 STP m3 CH4 kg
−1
VS added and a volumetric production of 1.23 STP m3 CH4 m
−3 day
−1 (Fig. 1a).
3.1.3. Food waste and card packaging
Over the first 100 days of operation the specific methane production (Fig. 2a) showed
an initial decline to a value between 0.26 and 0.28 STP m3 CH4 kg
−1 VSadded followed
by a peak and then a return to a value around 0.30 STP m3 CH4 kg
−1 VSadded. The biogas
methane content (Fig. 2b) showed some variability over this period in the range
50−60%, with an average around 54%. The VFA profile in the two digesters (Fig. 2c
and d) showed a very similar pattern, with the HPr concentration decreasing rapidly
after about 1 SRT coupled with increasing concentrations of HAc and, to a lesser extent,
of iso-valeric acid. The HAc was then consumed with a decrease from 5400 mg l−1
at
day 35 to 1100 mg l−1
at day 50, at which time HPr was less than 50 mg l−1
(Fig. 2c and
d). This pattern explains the changes in specific methane yield and biogas methane
concentration noted above. A further rise in HAc concentration was observed at the end
of the second SRT (day 50−60). Due to the low TKN content of card packaging, the
TAN fell sharply from an initial concentration of 3400 mg l−1
to 1800 mg l−1
by the end
of the second SRT, and this may have contributed to some extent to the recovery
process. During this period the free ammonia concentration dropped to less than 100 mg
l−1
(Fig. 2e) accompanied by a gradual fall in pH as the alkalinity dropped to <10,000
mg CaCO3 l−1
(Fig. 2f). The HAc concentration was stable at around 1500 mg l−1
during
the third SRT, although TAN continued to fall.
13
The OLR was increased from 2 to 3 kg VS m−3
day−1
at the beginning of the fourth SRT
time (day 99) which resulted in the volumetric methane production increasing from 0.60
to 0.69 m3 CH4 m
−3 day
−1 over the following 2 weeks (Fig. 2a). The VFA
concentrations were seen to increase shortly afterwards (Fig. 2c and d). With the
continuing fall in TAN (Fig. 2e) and alkalinity (Fig. 2f), the reduced buffering capacity
was insufficient to prevent a rapid decline in pH to a critical point (Fig. 2f). A hydrogen
peak was also noted in the GC analysis of biogas composition when the methane
content dropped to 25% and the digesters were failing. The VFA profile showed that the
condition of one digester (Fig. 2c) was worse than the other (Fig. 2d). In an attempt to
stabilise the pH and recover the most badly affected digester (FW + CP digester No. 1)
it was fed from day 134 with cattle slurry as the only substrate for 42 days, with the aim
of providing additional buffering and nutrients essential for anaerobic digestion and as a
source of methanogenic micro-organisms. The second digester of the pair (FW + CP
digester No. 2) was left without feeding for the same length of time. Both digesters
showed signs of recovery. The VFA concentration in the digester fed with cattle slurry
dropped from its peak value of 11,000 mg l−1
to a level of 1200 mg l−1
(Fig. 2c), the
biogas methane percentage reached 60% (Fig. 2b), and the pH rose to 7.3 (Fig. 2f). In
the other digester the pH climbed to around 7.5 (Fig. 2f), VFA were almost completely
consumed (Fig. 2d), and TAN rose (Fig. 2e) as degradation of the remaining organic
materials in the digestate increased the buffering capacity. After recovery both digesters
were again fed with the mixture of food waste and card packaging at an OLR of 2 kg
VS m−3
day−1
from day 176. The cattle slurry recovered digester showed good
operational stability within 2 weeks with the total VFA concentration falling then
14
remaining below 100 mg l−1
(Fig. 2c). The digester without cattle slurry addition
showed a fall in pH to less than 7.0 (Fig. 2f) and an increase in the VFA concentration
to 3000 mg l−1
(Fig. 2d), and it was necessary intermittently to stop feed addition over
the following two weekends to maintain operational stability. Continuous feeding was
then resumed, after which the digester operated stably for around 2 SRT and showed a
similar performance to the cattle slurry recovered digester. It is interesting to note that
the persistent HAc concentration of around 1500 mg l−1
observed during the first part of
the run at an OLR of 2 kg VS m−3
day−1
was undetectable after the re-acclimation at the
start of the second part (Fig. 2c and d).
An explanation for the observed behaviour is as follows: The digesters were started
using an inoculum of digestate from a food waste digester which was operating at a
TAN concentration of around 3400 mg l−1
and was likely to have some inhibition of
acetoclastic methanogenesis (Karakashev et al., 2006; Schnurer and Nordberg, 2008).
Although, the addition of card packaging lowered the TAN concentration it was evident
from the existence of an acetic acid plateau (Fig. 2c and d) that the population of
acetoclastic methanogens or the acetate oxidisation pathway was not entirely restored
within the first three SRT. The increase in OLR from 2 to 3 kg VS m−3
day−1
increased
the load on hydrogenotrophic and/or acetoclastic pathways, and when the methane
production route was unable to consume all the hydrogen produced, product-induced
feedback inhibition would have caused VFA build-up. With the TAN concentration
reducing to less than 1000 mg l−1
there was insufficient buffering capacity to prevent the
pH falling to a level at which the hydrogenotrophic methanogens were affected,
bringing the situation to a critical point. The recovery stage allowed the recovery of
15
methanogenesis in the digesters and developed a more balanced anaerobic microbial
consortium with conditions ideal for both acetoclastic and hydrogenotrophic
methanogenesis.
On day 232 the OLR in both digesters was again increased, to 3 kg VS m−3
day−1
. On
this attempt the VFA concentrations and all other operational parameters remained
stable (Figure 2b−f). The specific methane production remained the same at 0.315 STP
m3 CH4 kg
−1 VS added and the volumetric production increased by 50% (Fig. 2a). The
OLR was increased to 4 kg VS m−3
day−1
at the end of the tenth SRT (day 295). By the
end of the trial the digesters had been running at this loading for more than one SRT
without any fluctuation in stability parameters, and had apparently reached a steady
state condition, with a specific methane production of 0.307 STP m3 kg
−1 VS added and
a volumetric methane production of 1.23 STP m3 m
−3 day
−1 (Fig. 2a).
3.2. 75-litre digester trial
As noted above, the 4-litre digesters were initially inoculated with digestate from a food
waste digester that was showing signs of incipient instability. This was confirmed by
the subsequent failure of the food waste (FW)-only control digester, which continued
operating with the same feedstock and OLR. From two SRT onwards it was evident that
there were process benefits from adding cattle slurry as co-substrate, as this allowed an
increase in OLR from 2 to 4 kg VS m−3
day−1
, with the proportion of food waste in the
mix also being increased. The co-digestion of food waste and card packaging failed to
stabilise on the first attempt to raise the OLR from 2 to 3 kg VS m−3
day−1
, but the
second attempt was successful. This laboratory-scale trial, however, left unanswered
16
questions as to why the control digester had failed, and why the addition of cattle slurry
or card packaging had recovered an apparently failing inoculum.
To resolve some of these questions, three 75-litre digesters were set up, one of FW-only
and one on each of the two mixes, to see if the previous results were reproducible
starting with fresh inoculum from a municipal wastewater biosolids digester. The
characteristics of this inoculum, including concentrations of macro-nutrients and
potentially toxic elements (PTE), are shown in Table 3.
3.2.1. Co-digestion results
Experimental results from the 75-litre trial are shown graphically in Fig. 3. It can be
seen that methane production did not stabilise until after around 3 SRT (Fig. 3a). Before
this, two obvious peaks were observed in each of the digesters. The first appeared
towards the end of the first SRT (day 20), and occurred simultaneously in all three
digesters; this was followed by a fall in specific methane production, a lower biogas
methane content (Fig. 3b), a rapid accumulation of VFA (Fig. 3c), and a drop in pH
(Fig. 3e). As this initial production of VFA was consumed, a second biogas production
peak appeared. This happened first in the digester running on FW + CS (day 42), second
in the digester with FW as the sole substrate (day 49), and last in the one fed with FW +
CP (day 63). The second biogas production peaks also had an associated peak in biogas
methane content, as seen in Fig. 3b. After the second biogas peak this parameter
stabilised, although there were slight fluctuations later in the digester fed with the FW +
CS mix: these may have been due to differences between batches of cattle slurry
collected and used. This non-steady behaviour over the first 3 SRT is probably due to
17
the need for acclimation of the microbial consortium to the new feedstock and new
conditions in the digester. The more robust acid-forming bacteria appear to have
adapted to the substrate after the digesters had been running for around half a SRT, and
started to consume the accumulated feedstock at an accelerated rate, in the process
producing VFA faster than it could be consumed by the acetate oxidising bacteria and
methanogens. At this point perhaps the drop in pH provided optimal conditions for them
to consume the VFA, especially acetate; or the moderate OLR of 2 kg VS m−3
day−1
simply allowed the more sensitive methanogens to catch up with the acidogens before
the pH dropped to a critical point.
After 3 SRT the co-digestions with cattle slurry and card packaging showed a similar
specific methane production to that measured in the 4-litre trial at the same OLR of 2 kg
VS m−3
day−1
. All operational parameters were also in the safe range, with TAN less
than 2500 mg l−1
(Fig. 3d) and VFA below 100 mg l−1
(Fig. 3c) throughout the next 7
SRT (210 days) in the trial. Volumetric methane production was 0.43 and 0.64 STP m3
CH4 m−3
day−1
for FW + CS and FW + CP co-digestion respectively.
The difference in TAN observed between the 4-litre and 75-litre FW + CP co-digestion
trials was because the card packaging used in the smaller-scale trial was pre-processed
by wet maceration. The addition of water during this pre-treatment reduced the total
solids content of the feedstock from 94% to 20%, thus increasing the rate at which TAN
was hydraulically flushed out of the 4-litre digesters.
3.2.2. Food waste only
18
Although the digester fed on FW-only appeared to reach a steady state for biogas
production, the TAN concentration exceeded 4000 mg l−1
by day 150 and continued to
increase (Fig. 3d). By day 180 there were signs of stress, with an elevated VFA
concentration (Fig. 3c) from the previously stable level between day 60 and day 180 of
around 200 mg l−1
. Total VFA concentration built up rapidly to above 1000 mg l−1
by
day 210. Initially the VFA was predominantly in the form of HAc but later all species
were present, with an initial rise in HPr concentration. The concentration of HPr then
fell, accompanied by a further HAc accumulation (Fig. 3f). It can be seen that by the
end of the trial VFA concentrations had been fluctuating around 2000−4000 mg l−1
for
about three retention times. Although VFA concentrations were not as high as some
previously seen in digesters fed only on food waste (Banks et al., 2012; Banks et al.,
2008), the pattern of accumulation was similar. The pH in the digester remained high at
8.1 (Fig. 3e) due to sufficient buffering capacity provided by TAN giving an alkalinity
over 25,000 mg l−1
.
Concentrations of certain key trace elements in both the 4-litre and the 75-litre co-
digestion trials were analysed part way through the study as parallel work had shown
these were critical (Banks et al., 2012): the results are given in Table 4. At the time
when the trace element analysis was carried out, the 4-litre digestion trial had run for
161 days or 5.4 SRT, and the 75-litre trial for 140 days (4.7 SRT). Modelled dilute-out
profiles for essential trace elements (Fig. 4) suggested that concentrations of cobalt,
selenium and tungsten had all dropped to below 1 mg kg−1
TS in the 75-litre food waste
digester by day 210 when the VFA concentration had built up to above 1000 mg l−1
.
19
This corresponds to the concentration range at which other laboratory-scale food waste
co-digestion trials have also failed (Banks et al., 2012).
The results confirmed that the digestion of source segregated domestic food waste by
itself led to an increase in VFA, probably as a result of wash-out of essential trace
elements combined with an increasing TAN concentration. This condition was
prevented by co-digestion with cattle slurry or card packaging, which could contribute
essential elements and also reduce the ammonia concentration.
3.2.3. Digestate characteristics
The characteristics of the whole digestate, digestate liquor and digestate fibre from the
75-litre digesters are shown in Table 5. The FW-only digestate contained higher
concentrations of plant nutrients than digestate from the FW and CS co-digestion trials:
this is due not only to the high nitrogen, phosphorus and potassium (NPK) content in
food waste, but also because these elements are conserved in the digestion process
whilst the food waste input loses a higher proportion of its solids than the mixed
feedstock. Concentrations of plant nutrients and PTE in the digestate largely reflected
those in the corresponding substrate or mix. All of the samples met the UK standards for
PTE (PAS 110, 2010), confirming that co-digestion with CS is unlikely to reduce the
high quality of the digestate product obtained from source separated domestic food
waste.
Results from the digestate stability tests are shown in Fig. 5. Digestate fibre from the
FW + CS digester had a relatively high residual methane potential of 0.196 STP m3 kg
−1
20
VS (Fig. 5a); this was because cattle slurry contains a high proportion of lignocellulosic
materials which are only slowly degraded and, since the material also has a high water
content, it has a relatively short retention time in the digester. Digestate liquor from the
FW + CS digester had a lower residual methane potential (0.093 STP m3 kg
−1 VS, Fig.
5b) than the digestate fibre, perhaps due to the lignocellulosic materials in the fibre
component. The residual methane potentials from FW + CP digestate fibre (0.110 STP
m3 kg
−1 VS, Fig. 5c) and digestate liquor (0.079 STP m
3 kg
−1 VS, Fig. 5d) were less
than that from FW + CS digestate fibre and liquor. FW-only digestate had the highest
residual methane potential of 0.203 STP m3 kg
−1 VS (Fig. 5e); this was probably
because the larger-scale food waste digester had VFA concentration of 2600 mg l−1
when the digestate sample was taken on day 286. The UK PAS 110 (2010) sets
digestate stability limits of VFA <0.43 g COD g−1
VS and residual biogas potential
<0.25 l g−1
VS in a 28-day test duration. According to these criteria, both fractions of
the mixed FW + CS and FW + CP digestate would pass. This requirement may however
suggest a reason for not uncoupling the solids and liquids retention time, as the
relatively short SRT of 30 days means a higher proportion of the digestate solids will
not be broken down. Although the FW-only digestate also passed the UK PAS 110
(2010) digestate stability limits, it would probably fail if the digester was run for a
longer period and the expected gradual build-up of VFA continued (Banks et al., 2012,
2011b, 2008).
4. Conclusions
21
Digestion of source-segregated domestic food waste at low OLR showed long-term
accumulation of volatile fatty acids, confirming results from earlier work. Co-digestion
with cattle slurry and card packaging reduced ammonia concentrations to non-inhibitory
levels, and allowed an increase in OLR to 4 kg VS m−3
day−1
with enhanced volumetric
methane productions. Both co-digestates were low in potentially toxic elements and
showed low residual methane potential. Trace element analysis suggested food waste
may lack cobalt, selenium and tungsten and these may be supplemented by the co-
digestates. Pilot-scale studies demonstrated long-term process stability of co-digestion
with cattle slurry and card packaging compared to mono-digestion of food waste.
Acknowledgements
The authors wish to thank the UK Government's Department for Environment Food and
Rural Affairs (Defra) and the EU 7th Framework programme for supporting this work
through Grant Nos. WR1208 and 241334 (VALORGAS), respectively.
References
1. APHA, 2005. Standard Methods for the Examination of Water and Wastewater, 21st
ed. American Public Health Association, American Water Works Association, Water
Environment Federation, Washington, USA.
2. Ahring, B.K., 2003. Perspectives for anaerobic digestion. Adv. Biochem. Eng.
Biotechnol. 81, 1-30.
3. Alvarez, R., Lidén, G., 2008. Semi-continuous co-digestion of solid slaughterhouse
waste, manure, and fruit and vegetable waste. Renew. Energ. 33, 726-734.
22
4. Angelidaki, I., Ellegaard, L., 2003. Codigestion of manure and organic wastes in
centralized biogas plants. Appl. Biochem. Biotechnol. 109, 95–105.
5. Banks, C.J., Zhang, Y., Jiang, Y., Heaven, S., 2012. Trace element requirements for
stable food waste digestion at elevated ammonia concentrations, Bioresour. Technol.
104, 127-135.
6. Banks, C.J., Salter, A.M., Heaven, S., Riley, K., 2011a. Energetic and environmental
benefits of co-digestion of food waste and cattle slurry: A preliminary assessment.
Resour. Conserv. Recy. 56, 71-79.
7. Banks, C.J., Chesshire, M., Heaven, S., Arnold, R., 2011b. Anaerobic digestion of
source-segregated domestic food waste: Performance assessment by mass and energy
balance. Bioresour. Technol. 102, 612-620.
8. Banks, C.J., Zhang, Y., 2010. Technical Report: Optimising inputs and outputs from
anaerobic digestion processes. Defra Project Code WR0212. Department of the
Environment, Food and Rural Affairs, London.
9. Banks, C.J., Chesshire, M., Stringfellow, A., 2008. A pilot-scale comparison of
mesophilic and thermophilic digestion of source segregated domestic food waste.
Water Sci. Technol. 58, 1475-1480.
10. Banks, C.J., Salter, A.M., Chesshire, M., 2007. Potential of anaerobic digestion for
mitigation of greenhouse gas emissions and production of renewable energy from
agriculture: barriers and incentives to widespread adoption in Europe. Water Sci.
Technol. 55, 165-173.
11. BioAbV, 1998. Verordnung über die Verwertung von Bioabfällen auf
landwirtschaftlich, forstwirtschaftlich und gärtnerisch genutzten Böden
23
(Bioabfallverordnung – BioAbfV). Bundesministerium der Justiz, Germany.
http://www.gesetze-im-internet.de/bioabfv/, accessed on 20th December 2011.
12. Braun, R., Wellinger, A., 2003. Potential of Co-digestion. IEA Task 37.
http://www.iea-biogas.net/_download/publi-
task37/Potential%20of%20Codigestion%20short%20Brosch221203.pdf, accessed on
20th December 2011.
13. Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F., 2002. Continuous co-
digestion of cattle slurry with fruit and vegetable wastes and chicken manure.
Biomass Bioenerg. 22, 71-77.
14. Capela, I., Rodrigues, A., Silva, F., Nadais, H., Arroja, L., 2007. Impact of industrial
sludge and cattle manure on anaerobic digestion of the OFMSW under mesophilic
conditions. Biomass Bioenerg. 32, 245-251.
15. Clemens, J., Trimborn, M., Weiland, P., Amon, B., 2006. Mitigation of Greenhouse
Gas Emissions from Livestock Production. Agr. Ecosyst. Environ. 112, 171-177.
16. El-Mashad, H.M., Zhang, R., 2010. Biogas production from co-digestion of dairy
manure and food waste. Bioresour. Technol. 101, 4021-4028.
17. Hartmann, H., Ahring, B.K., 2005. Anaerobic digestion of the organic fraction of
municipal solid waste: Influence of co-digestion with manure. Water Res. 39, 1543-
1552.
18. Holm-Nielsen, J.B., Al Seadi, T., Oleskowicz-Popiel, P., 2009. The future of
anaerobic digestion and biogas utilization. Bioresour. Technol. 100 ,5478-5484.
19. Karakashev, D., Batstone, D.J., Trably, E., Angelidaki, I., 2006. Acetate oxidation is
the dominant methanogenic pathway from acetate in the absence of
Methanosaetaceae. Appl. Environ. Microbiol. 72, 5138-5141.
24
20. Lukehurst, C.T., Frost, P., Al Seadi, T., 2010. Utilisation of digestate from biogas
plants as biofertiliser. IEA Task 37. http://www.iea-biogas.net/_download/publi-
task37/Task37_Digestate_brochure9-2010.pdf, accessed on 20th December 2011.
21. Marañón, E., Salter, A.M., Castrillón, L., Heaven, S., Fernández-Nava, Y., 2011.
Reducing the environmental impact of methane emissions from dairy farms by
anaerobic digestion of cattle waste. Waste Manage. 31, 1745-1751.
22. Mata-Alvarez, J. (Ed.), 2003. Biomethanisation of the organic fraction of municipal
solid wastes. IWA Publishing, London.
23. Neiva Correia, C., Vaz, F., Torres, A., 2008. Anaerobic Digestion of Biodegradable
Waste – Operational and Stability Parameters for Stability Control. In: 5th IWA Int.
Symp. on AD of Solid Wastes & Energy Crops, Tunisia.
24. Nielsen, H.B., Angelidaki, I., 2008. Codigestion of manure and organic waste at
centralized biogas plants:process imbalances and limitations. Water Sci. Technol. 58,
1521-1528.
25. Palatsi, J., Viñas, M., Guivernau, M., Fernandez, B., Flotats, X., 2011. Anaerobic
digestion of slaughterhouse waste: Main process limitations and microbial
community interactions. Bioresour. Technol. 102, 2219-2227.
26. PAS 110, 2010. Specification for whole digestate, separated liquor and separated
fibre derived from the anaerobic digestion of source-segregated biodegradable
materials. British Standards Institution, London.
27. Raven, R.P.J.M., Gregersen, K.H., 2007. Biogas plants in Denmark: successes and
setbacks. Renew. Sust. Energ. Rev. 11, 116–132.
25
28. Resch, C., Worl, A., Waltenberger, R., Braun R., Kirchmayr, R., 2011.
Enhancement options for the utilisation of nitrogen rich animal by-products in
anaerobic digestion. Bioresour. Technol. 102, 2503-2510.
29. Ripley, L.E., Boyle, W.C., Converse, J.C., 1986. Improved alkalimetric monitoring
for anaerobic digestion of high-strength wastes. J. Water Pollut. Control Fed. 58,
406-411.
30. SCA, 1986. Methods for the determination of metals in soils, sediments and sewage
sludge and plants by hydrochloric-nitric acid digestion. Methods for the Examination
of Waters and Associated Materials, Standing Committee of Analysts, HMSO,
London.
31. Schnurer, A., Nordberg, A., 2008. Ammonia, a selective agent for methane
production by syntrophic acetate oxidation at mesophilic temperature. Water Sci.
Technol. 57, 735-740.
32. Walker, M., Zhang, Y., Heaven, S., Banks, C.J., 2009. Potential errors in the
quantitative evaluation of biogas production in anaerobic digestion processes.
Bioresour. Technol. 100, 6339-6346.
33. Zhang, L., Lee, Y.-W., Jahng, D., 2011. Anaerobic co-digestion of food waste and
piggery wastewater: Focusing on the role of trace elements. Bioresour. Technol. 102,
5048-5059.
26
Figure captions
Fig. 1. Operational performance and stability characteristics of FW and CS co-digestion
at 4-litre scale. Vertical dashed lines indicate when a change of OLR or VS ratio of FW
to CS took place.
Fig. 2. Operational performance and stability characteristics of FW and CP co-digestion
at 4-litre scale. Vertical dashed lines indicate when a change of OLR or feeding regime
took place.
Fig. 3. Operational performance and stability characteristics of FW and co-substrates
digestion at 75-litre scale.
Fig. 4. Simulated essential elements concentration-time profile in the 75-litre food
waste only digester (day 0 and 140 was when a sample of digestate was taken for trace
element analysis and day 210 was when VFA level rose to above 1000 mg l-1
.)
Fig. 5. Residual methane potential test on digestate from the 75-litre digestion trial.
27
Table 1
Substrate characteristics.
Food waste Cattle slurry Card packagingd
General
pH (1:5 dilution for FW and CS, 1:30 for CP) 4.71 ± 0.01 7.83 ± 0.07 7.21± 0.03
TS (% wet weight (WW)) 23.74 ± 0.08 9.31 ± 0.14 93.9 ± 0.1
VS (% WW) 21.71 ± 0.09 6.52 ± 0.04 78.5 ± 0.4
VS (% of TS) 91.44 ± 0.39 70.0 ± 0.6 83.6 ± 0.5
Total Organic Carbon (TOC) (% of TS) 47.6 ± 0.5 38.9 ± 1.0 41.6 ± 0.7
TOC / TKN 13.9 ± 0.2 11.1 ± 0.3 288 ± 5
Biodegradable C a / TKN
13.6 ± 0.2 8.15 ± 0.32 42.9 ± 4.7
Calorific value (CV) (kJ g-1
TS) 20.7 ± 0.2 16.75±0.10 17.18 ± 0.36
Biochemical composition (VS basis)
Carbohydrates b (g kg
-1) 453 ± 17 21.9 ± 1.0 242 ± 19
Lipids c (g kg
-1) 151 ± 1 93.6 ± 0.8 < 10
Crude proteins (g kg-1
) 235 ± 3 276 ± 6 10.8 ± 0.0
Hemi-cellulose (g kg-1
) 38.1 ± 3.7 226 ± 6 113 ± 5
Cellulose (g kg-1
) 50.4 ± 1.6 96.7 ± 8.5 304 ± 6
Lignin (g kg-1
) 16.5 ± 0.2 226 ± 7 532 ± 2
NPK and PTE content (TS basis)
Total Kjeldahl nitrogen (TKN) (g kg-1
) 34.2 ± 0.4 35.0 ± 0.5 1.44 ± 0.01
Total Phosphorus (TP) (g kg-1
) 5.41 ± 0.32 8.58 ± 0.63 0.134 ± 0.003
Total Potassium (TK) (g kg-1
) 14.3 ± 0.8 16.7 ± 0.2 0.221 ± 0.010
Cd (mg kg-1
) < 1.0 < 1.0 < 0.05
Cr (mg kg-1
) 29.0 ± 1.2 113 ± 2 9.1 ± 0.9
Cu (mg kg-1
) 7.20 ± 0.81 58.4 ± 1.1 20.3 ± 2.3
Hg (mg kg-1
) < 0.010 < 0.010 < 0.10
Ni (mg kg-1
) 7.0 ± 2.9 44.8 ± 0.6 4.5 ± 0.5
Pb (mg kg-1
) < 10 < 10 2.9 ± 0.4
Zn (mg kg-1
) 33 ± 11 231 ± 6 16.2 ± 4.3
Elemental composition (TS basis)
N (%) 3.44 ± 0.04 3.50 ± 0.05 0.14± 0.00
C (%) 47.6 ± 0.5 38.9 ± 1.0 41.6 ± 0.7
H (%) 7.04 ± 0.63 5.18 ± 0.15 4.76 ± 0.23
S (%) 0.15 ± 0.01 0.31 ± 0.02 0.21 ± 0.00
O (%) 33.3 ± 2.6 23.1 ± 0.9 36.9 ± 0.9 a Biodegradable carbon was calculated by deducting lignin carbon from TOC. The formula of lignin was
chosen as C9H7.95O2.41(OMe)0.93; b In equivalent glucose;
c n-hexane extractable material (HEM);
d Card packaging used for characteristics determination was the dry material as used in the 75-litre trial.
28
Table 2
Operational regimes for the 4-litre laboratory-scale digestion trials.
Time (days) FW+CS1 FW+CS2 FW+CP1 FW+CP2 FW
35
OLR = 2,
FW:CS =
20:80a
OLR = 2,
FW:CS =
20:80
OLR = 2b
OLR = 2
OLR = 2
98
Ceased
feeding from
day 35
onwards
133 OLR = 2,
FW:CS =
40:60
OLR = 2,
FW:CS =
40:60
OLR = 3 OLR = 3
175 Fed with cattle
slurry
Ceased
feeding
231
OLR = 3,
FW:CS =
40:60
OLR = 3,
FW:CS =
40:60
OLR = 2 OLR = 2c
294
OLR = 3,
FW:CS =
60:40
OLR = 3,
FW:CS =
60:40
OLR=3 OLR=3
329
OLR = 4,
FW:CS =
60:40
OLR = 4,
FW:CS =
60:40
OLR = 4 OLR = 4
Note: FW – Food Waste; CS – Cattle Slurry; CP – Card packaging a The proportion of food waste and cattle slurry was on a VS basis;
b The proportion of food waste and card packaging was 53:47 on a VS basis throughout the 4-litre trial;
c No feeding at the first two weekends.
29
Table 3
Nutrient (NPK), PTE and solids content of the wastewater biosolids digestate used as
the inoculum for the 75-litre trial.
Whole digestate Liquor fraction Fibre fraction
TAN (g NH3-N kg-1
TS) 39.1 ± 0.5 40.1 ± 0.3 -
TKN (g N kg-1
TS) 77.5 ± 1.5 79.3 ± 1.1 43.4 ± 1.0
TK (g K kg-1
TS) 2.90 ± 0.29 2.99 ± 0.13 1.26 ± 0.26
TP (g P kg-1
TS) 32.4 ± 3.5 33.2 ± 1.3 16.8 ± 2.6
Cd (mg kg-1
TS) 1.10 ± 0.21 1.10 ± 0.10 < 1.0
Cr (mg kg-1
TS) 67.3 ± 5.3 68.8 ± 5.3 40.1 ± 2.2
Cu (mg kg-1
TS) 462 ± 9 473 ± 1 247 ± 24
Ni (mg kg-1
TS) 52.9 ± 7.4 54.0 ± 7.5 32.5 ± 2.1
Pb (mg kg-1
TS) 83.8 ± 8.4 83.9 ± 0.7 61.9 ± 8.0
Zn (mg kg-1
TS) 718 ± 27 736 ± 24 380 ± 24
TS (% WW) 4.48 ± 0.07 4.10 ± 0.01 10.8 ± 0.6
VS (% WW) 2.81 ± 0.02 2.54 ± 0.01 8.56 ± 0.58
VS (% TS) 62.8 ± 1.4 62.0 ± 0.1 79.8 ± 0.5
30
Table 4
Concentrations of essential elements (mg kg-1
TS) in whole digestate samples from 4-
litre and 75-litre co-digestion trials.
Digester Co Cu Mn Mo Ni Se W Zn
4-litre food waste co-digestion trials (day 161)a
FW + CS, No. 1 2.0 63.2 667 6.9 62.4 0.46 3.1 241
FW + CP, No. 1 1.2 75.7 443 3.3 28.4 0.26 < 0.25 173
FW + CP, No. 2 1.2 73.9 171 5.9 53.5 0.16 3.3 96.8
FW control 1.3 140 255 7.9 132 0.33 6.1 121
75-litre food waste co-digestion trials (day 140)
FW + CS 1.8 196 696 3.9 36.3 0.59 < 0.25 314
FW + CP 1.3 178 163 3.3 36.5 < 0.15 2.1 198
FW control 1.5 209 367 3.1 65.0 0.57 1.6 287 a Samples for trace element analysis were taken from FW + CS digester No. 1, and from both FW + CP
digesters to take account of the different treatments (FP + CP No. 1 fed with CS from Day 134 - 176, FW
+ CP No. 2 not fed for the same period).
31
Table 5
Characteristics of whole digestate, digestate liquor and digestate fibre in 75-litre trials.
Whole digestate Digestate liquor Digestate fibre
Bio
AbV
PAS
110
FW +
CS
FW +
CP
FW
only
FW +
CS
FW +
CP
FW
only
FW +
CS
FW +
CP
FW
only
Nutrients (g kg-1
TS)
TKN 58.8 65.3 136 65.2 70.6 138 32.2 33 62.2
TP 13 8.9 14 14 10 14 5.6 3.1 13
TK 34 25 50 38 29 50 11 9.5 24
Potentially toxic elements (mg kg-1
TS)
Cd <0.50 <0.50 <0.50 <1.0 <1.0 <1.0 <0.50 <0.50 <0.50 1.0 1.5
Cr 31 53 79 23 60 79 68 24 51 70 100
Cu 100 160 130 110 180 130 46 61 71 70 200
Pb <10 52 18 <18 62 18 <10 14 <10 100 200
Hg <0.25 <0.25 <0.25 <0.50 <0.50 <0.50 <0.25 <0.25 <0.25 0.7 1
Ni 16 29 48 13 33 49 32 13 28 35 50
Zn 270 130 180 300 150 180 140 50 130 300 400
Essential elements (mg kg-1
TS)
Co <1.0 <1.0 <1.0 <2.0 <2.0 <2.0 1.3 <1.0 <1.0
Fe <2000 3200 5700 <4000 4100 5700 <2000 <2000 2700
Mo 3.9 5.4 5.2 4.2 6.2 5.2 2.3 2.3 2.7
Se <0.30 <0.30 0.7 <0.50 <0.50 0.71 <0.30 <0.30 <0.30
W <1.0 <1.0 <1.0 <2.0 <2.0 <2.0 <1.0 <1.0 <1.0
Solids content (% WW)
TS 5.84 6.76 6.04 4.72 5.16 5.95 11.4 12.7 14.1
VS 4.38 4.45 4.19 3.34 3.21 4.1 9.7 9.33 11.2
32
Fig. 1.
33
Fig. 2.
34
Fig. 3.
35
Fig. 4.
36
Fig. 5.